Design And Implementation Of Text Affective Computing System Based On Dictionary

With the delvelopment of the Internet and the rapid advancement of mobile Internet,nowadays,net has became a very important stage for people to express their views andpoints. Coupled with the birth and development of a variety of internet products,makingthe network public opinion occipies an important position in the public opinion, the studyand the discuess of public opinion are get more and more attention. Belong the researchesin the public opinion of network,the important point is that sentiment analysis.Conductingsentiment analysis of public opinion can qucikly find and dig out the users’ attitude aboutsome time,and futher make reliable predictions about the development of events,also canfliter out the public opinion influence brought by the "water army" to a certain extent.Sentiment Analysis is mainly taken from the text to determine the emotionaltendencies that are positive, neutral and negative. According to the characteristics of thepractical application of this article,based on the existing basis of the traditionalcalculation,for different lengths of the text using the different emotion analysismethods.For long text classfication proposed to build a theme dictionary assistedmethod, using the found of new words and to be artificial auxiliary construct a themedictionary.Long text sentiment analysis, classification and then use the themedictionaries auxiliary analysis of emotions. For essay directly using related emotionaldictionary polarity weights calculation methods to get emotional tendencies.Throughemotional dictionary building and polarity weights of fine-tuning, make the result moreaccurate and reliable.For long text, with actual use of the three gorges project of thenews text sentiment analysis, for short text,using data sina weibo for emotionalcalculated,both had achieved ideal result.